Caching in docker implemented

Dockerfile

@@ -8,14 +8,32 @@ WORKDIR /app
 COPY . /app
 
 # Install any needed packages specified in requirements.txt
-RUN pip install --trusted-host pypi.python.org -r requirements.txt
-    pip uninstall transformers
+RUN pip install --trusted-host pypi.python.org -r requirements.txt && \
+    pip uninstall transformers && \
     pip install transformers==4.29.2
 
 # Make port 80 available to the world outside this container
 EXPOSE 80
 
-RUN ls -al
+# Set the TORTOISE_MODELS_DIR environment variable
+ENV TORTOISE_MODELS_DIR tortoise/models/pretrained_models
+
+# Create the directory for pretrained models
+RUN mkdir -p $TORTOISE_MODELS_DIR
+
+# Download all the models
+RUN wget -O $TORTOISE_MODELS_DIR/autoregressive.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/autoregressive.pth && \
+    wget -O $TORTOISE_MODELS_DIR/classifier.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/classifier.pth && \
+    wget -O $TORTOISE_MODELS_DIR/clvp2.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/clvp2.pth && \
+    wget -O $TORTOISE_MODELS_DIR/cvvp.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/cvvp.pth && \
+    wget -O $TORTOISE_MODELS_DIR/diffusion_decoder.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/diffusion_decoder.pth && \
+    wget -O $TORTOISE_MODELS_DIR/vocoder.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/vocoder.pth && \
+    wget -O $TORTOISE_MODELS_DIR/rlg_auto.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/rlg_auto.pth && \
+    wget -O $TORTOISE_MODELS_DIR/rlg_diffuser.pth https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/rlg_diffuser.pth && \
+    wget -O $TORTOISE_MODELS_DIR/bigvgan_base_24khz_100band_g.pth https://drive.google.com/uc?id=1_cKskUDuvxQJUEBwdgjAxKuDTUW6kPdY && \
+    wget -O $TORTOISE_MODELS_DIR/bigvgan_24khz_100band_g.pth https://drive.google.com/uc?id=1wmP_mAs7d00KHVfVEl8B5Gb72Kzpcavp
+
+RUN ls -la /app
 
 # Run app.py when the container launches
 CMD ["streamlit","run", "app.py"]

app.py

@@ -15,6 +15,9 @@ PRESETS = ["ultra_fast", "fast", "standard", "high_quality", "very_fast"]
 UPLOAD_FOLDER = "./uploads"
 OUTPUT_FOLDER = "./output"
 
+voice_samples = None
+conditioning_latents = None
+
 # Create upload and output directories if they don't exist
 os.makedirs(UPLOAD_FOLDER, exist_ok=True)
 os.makedirs(OUTPUT_FOLDER, exist_ok=True)
@@ -46,6 +49,7 @@ if st.sidebar.button("Create Voice") and voice_name.strip() != "":
             wav_file.write(bytes_data)
 
     st.sidebar.success(f"Voice '{voice_name}' created successfully!")
+    voice_samples, conditioning_latents = load_voice(voice_name)
 
 # Input text and settings
 st.header("Text-to-Speech Generation")
@@ -68,7 +72,9 @@ if st.button("Generate Speech"):
         st.info("Generating speech...")
 
         # Load voice samples
-        voice_samples, conditioning_latents = load_voice(voice)
+        # voice_samples, conditioning_latents = load_voice(voice)
+        
+        print(voice_samples)
 
         # Generate speech with Tortoise
         gen = tts.tts_with_preset(
@@ -82,6 +88,8 @@ if st.button("Generate Speech"):
         output_path = os.path.join(OUTPUT_FOLDER, "generated.wav")
         torchaudio.save(output_path, gen.squeeze(0).cpu(), 24000)
 
+        print(output_path)
+
         # Log the path of the generated audio
         st.write(f"Generated audio saved at: {output_path}")
 

tortoise/models/utils.py

@@ -11,10 +11,12 @@ import progressbar
 
 D_STEM = "https://drive.google.com/uc?id="
 
-DEFAULT_MODELS_DIR = os.path.join(
-    os.path.expanduser("~"), ".cache", "tortoise", "models"
-)
-MODELS_DIR = os.environ.get("TORTOISE_MODELS_DIR", DEFAULT_MODELS_DIR)
+# DEFAULT_MODELS_DIR = os.path.join(
+#     os.path.expanduser("~"), ".cache", "tortoise", "models"
+# )
+# MODELS_DIR = os.environ.get("TORTOISE_MODELS_DIR", DEFAULT_MODELS_DIR)
+
+MODELS_DIR = os.environ.get("TORTOISE_MODELS_DIR")
 MODELS = {
     "autoregressive.pth": "https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/autoregressive.pth",
     "classifier.pth": "https://huggingface.co/jbetker/tortoise-tts-v2/resolve/main/.models/classifier.pth",
@@ -76,5 +78,5 @@ def get_model_path(model_name, models_dir=MODELS_DIR):
     return model_path
 
 if __name__ == "__main__":
-    download_models() # to download all models
+    # download_models() # to download all models